In some cases, you may have existing automated analyses (e.g. already kicked off as part of existing CI pipeline), or results generated by local, internal, or remote processes not being kicked off directly by Miqa. To run automated tests on these existing “offline” outputs, they must be generated in a cloud location accessible to your Miqa processor, and your pipeline must notify Miqa of this location.
You can notify Miqa of the output cloud location of each dataset, and trigger batch testing, via REST API, Python tools, or CSV upload.
CSV Upload
Batch processing by CSV is a very convenient and powerful way to add many datasets and/or new results at one time. You can quickly point Miqa to results for one, dozens, or hundreds of samples across one or multiple pipeline versions, if you have those results available from executing offline. Miqa can then perform testing across these results on each specified version.
Start by preparing a CSV file that looks like this:
For each new offline-upload version to add, use a column header prefixed by "newversion_", e.g. "newversion_VERSIONNAME". If you are adding more dataset results for a version already registered with Miqa, instead use the prefix "existingversion_", e.g. "existingversion_VERSIONNAME".
If adding new datasets, include a "name" column with the dataset names, and in the "id" column use the text "NEW". This will signal to Miqa to create a new Dataset within the system, corresponding to the name provided.
Populate the fields in C2 onward (blue highlighted section) with the cloud location for the outputs that correspond to that row's dataset and that column's pipeline version.
REST API
To batch process results from a script or other trigger, we recommend using either the Python tools or REST API. This requires initial setup of a test trigger that specifies the datasets that are included in the test. Your onboarding/support representative will guide you to create your first trigger.
Once set up, you will use the Trigger Test & Set Output Details (Offline Results) API endpoint to pass your trigger, version, and output locations.
